Is "development length" more popular than "anchorage length" in the US?

[asiga]

Coming from Europe, I'm very used to the term "anchorage length" (and in fact I've never seen it called "development length" in my day to day environment). Now, reading some RC books published in the US, I see they tend to prefer "development length" (and... doing an Internet search, I even found some posts arguing they are different concepts --I assume they are wrong, in fact their argument doesn't make any sense).

Would "anchorage length" be understood in the US? Or would it sound weird?

Do you in the US use "anchorage" just for the concept, but not for the length used in design?

Thanks!

 

[KootK]

doing an Internet search, I even found some posts arguing they are different concepts --I assume they are wrong, in fact their argument doesn't make any sense

Yeah, I'm your huckleberry. Please tell me how you feel that "their" argument doesn't make any sense so that I can begin the evisceration process in earnest.

Would "anchorage length" be understood in the US? Or would it sound weird?

I suspect that it would be understood most of the time but cause confusion occasionally now that we have concrete code provisions that deal with the issue of anchorage of things separate from development. I'm pretty sure that we still have some vestigial code clauses that treat anchorage and development of reinforcement interchangeably.

Do you in the US use "anchorage" just for the concept, but not for the length used in design?

I see the term "development length" used for:

1) overwhelmingly the concept that actually is development and;
2) often the concept of anchorage which, obviously, you deny the existence of as a separate concern.

 

[KootK]

In all seriousness though, I encourage a diversity of opinion and always welcome a debate on this particular topic.

 

[asiga]

the concept of anchorage which, obviously, you deny the existence of as a separate concern

In the EC-2, the anchorage length is the length that a reinforcement needs for the stress to be transferred safely to concrete (ie: without failure of the bond). There's no "development length" in the EC-2. And, when I read its meaning in US books, it's the same meaning as the anchorage length in the EC-2, so... what's the difference?

People arguing it's different seem to say that the development length takes place in straight-in cuts of rebars, while anchorage happens at ends of members, with bends and hooks (as I said, this sounds nonsense to me, because they are both the same situation: the need for extending the length of a rebar for safely transferring the stress is the same no matter if you are cutting curtailed reinforcements, or if you at at the ends of a member).

 

[asiga]

Oh, maybe you call it "development" when it's about a straight rebar, and "anchorage" when you use a standard bend/hook? If that's the case, I can understand your will to argue they are different, but we have always considered both within a homogeneous formulation (with coefficients that penalize straight anchorages against the curved ones, or the ones with welded transverse bars).

 

[KootK]

Firstly, I'm only loosely familiar with Eurocode presentation on anchorage/development so you may have to assist me by pointing me to what I might need to see.

In the EC-2, the anchorage length is the length that a reinforcement needs for the stress to be transferred safely to concrete (ie: without failure of the bond)...so... what's the difference?

Let's parse that.

...without failure of the bond

1) This is precisely what development means in North American codes. No radial splitting and the bar doesn't slide out of the cylindrical hole into which it is embedded prior to reaching its yield strength. This does not preclude the possible outcome that the reinforcement rips from the parent concrete, taking a chunky cone of concrete with it even though a pure bond stress failure has been precluded.

...the length that a reinforcement needs for the stress to be transferred safely to concrete...

2) In North America, safely transferring reinforcement tension to concrete would involve two, distinct aspects:

a) preventing a bond stress failure as described in #1 and conventionally termed "development length" in North America and;

b) preventing the reinforcement from ripping from the parent concrete without a bond stress failure by taking a chunk of concrete along with it. This is what is termed, or should be termed "anchorage" in North American codes.

People arguing it's different seem to say that the development length takes place in straight-in cuts of rebars, while anchorage happens at ends of members, with bends and hooks

No, that's a misrepresentation of what has been said. It would be more accurate to articulate the perspective as:

1) Development / bond stress limitation is required everywhere bar force is to be transferred out of a bar and into something else, whether that something else is nearby reinforcing or nearby concrete.

2) When bar tension is transferred from one group of reinforcing to another via a code compliant lap splice, no separate check for anchorage is required.

3) When bar tension is transferred from a group of reinforcing to a concrete mass, one of the following is required in addition to development:

a) Anchorage of the reinforcing within a concrete compression strut (most common) or;

b) Anchorage of the reinforcing via embedment into a concrete mass and utilizing the tension capacity of concrete (Eligenhausen stuff).

as I said, this sounds nonsense to me, because they are both the same situation: the need for extending the length of a rebar for safely transferring the stress is the same no matter if you are cutting curtailed reinforcements, or if you at at the ends of a member

Correct, in both of those examples, you would require both of the following, in the North American parlance:

4) Development and;

5) Anchorage. In both of the examples that you mentioned, the anchorage would commonly be the 3a type described above rather than the 3b/Eligenhausen type.

 

[KootK]

Oh, maybe you call it "development" when it's about a straight rebar, and "anchorage" when you use a standard bend/hook?

Nope. Both straight development and hooked development are just two different forms of the same thing that we call "development". Obviously the two details distribute bar force to the concrete differently. However, this is not something that is given explicit consideration in member design from any perspective other than intelligent detailing (which way to point the hook etc).

 

[asiga]

I understand, although I feel a bit surprised about how things are treated differently at each side of the Atlantic.

I think the key point is how much concrete thickness are you talking about when you say "the reinforcement rips from the parent concrete, taking a chunky cone of concrete with it even though a pure bond stress failure has been precluded"... if you mean a thickness no larger than the minimum allowed separation between rebars (ie: say 2-3 cm), then that mechanism of failure is (supposedly) included in the anchorage formulation in Europe (I say "supposedly" because that's what we learnt in our books here: there are small struts of compressed concrete starting out from the ribs of the rebar, and if such struts fail, the pull-out takes place with a thickness of concrete still surrounding the rebar).

These struts from the ribs are presented in our books as one of the three phenomena behind the bond between rebars and concrete, and are implicitly included in our "anchorage length" (because if they fail, it's because you are transferring too high a stress to the concrete, so your anchorage length is too short).

I said "if you mean a thickness no larger than the minimum allowed separation between rebars". Now, if you mean a larger thickness (say, 10cm, 15cm...), then we are talking about a different scenario, more related to concentrated loads (ie: supports for prestressed steel, or point loads on mass concrete volumes, etc), and here we would address this as with a strut-tie design, not with the anchorage length.

 

[asiga]

Perhaps the reason for these different approaches is that, for us (Europe), bond is made of three causes:
1) molecular forces (adhesion),
2) mechanical friction at the steel/concrete interface,
3) mechanical interlock in the compressed struts at the ribs.

If I'm understanding you correctly, you are considering 3) outside of the bond. This might be the reason why you have a "development length" and an "anchorage length". We take all the three in the bond, and so we just have "anchorage length".

By the way... do you mean you cannot break the struts in a lap splice? Sure you can! (you can make a lap splice fail with concrete still surrounding the rebars)

 

[KootK]

if you mean a thickness no larger than the minimum allowed separation between rebars (ie: say 2-3 cm), then that mechanism of failure is (supposedly) included in the anchorage formulation in Europe...

No, that is not what I mean. I would consider all of that to be encompassed within the development length concept.

(I say "supposedly" because that's what we learnt in our books here: there are small struts of compressed concrete starting out from the ribs of the rebar, and if such struts fail, the pull-out takes place with a thickness of concrete still surrounding the rebar).

I don't believe that is correct. Everything that I've seen suggests that localized splitting is the failure most correlated with the compression struts that emanate from the reinforcing bar deformations. Although, under testing, it is often bar slip that defines "failure" when cover and spacing are adequate to preclude splitting.

Now, if you mean a larger thickness (say, 10cm, 15cm...), then we are talking about a different scenario, more related to concentrated loads (ie: supports for prestressed steel, or point loads on mass concrete volumes, etc), and here we would address this as with a strut-tie design, not with the anchorage length.

I disagree. Reinforcing anchorage certainly tends to be more critical at disturbed regions often design with strut and tie. That said, reinforcing bar anchorage is a requirement in all situations where reinforcing bar tension is mobilized, including scenarios typicality designed via traditional, sectional methods. I've provided an example of this below for the case of cut-off bars which you mentioned previously. This is a situation where strut and tie methods would not commonly be employed and, rather, traditional sectional design methodologies would typically be used. Still, however, some form of reinforcing bar anchorage is required in addition to basic bar development. In this case, it is the 3a type of anchorage that I described previously and the "check" for this would be embodied in the conventional flexure and shear design checks.

 

[KootK]

Perhaps the reason for these different approaches is that, for us (Europe), bond is made of three causes:
1) molecular forces (adhesion),
2) mechanical friction at the steel/concrete interface,
3) mechanical interlock in the compressed struts at the ribs.

All three of those things are encompassed in the North American conception of development length so I don't see that as being the issue.

If I'm understanding you correctly, you are considering 3) outside of the bond.

Not at all. I consider #3 to be encompassed within the concept of development length.

By the way... do you mean you cannot break the struts in a lap splice?

No, that's not what I meant. Rather, what I meant was that a code compliant lap splice does not require any additional anchorage checks to verify that the little concrete struts that develop between the lapped bars are adequate. The capacity of those struts is already built into the procedure that is used to calculate splice length. It's "baked into the cake" so to speak.

Sure you can! (you can make a lap splice fail with concrete still surrounding the rebars)

I don't understand sure what you're getting at here. Can you provide a sketch or a photo of the failure mode that you're envisioning? In a code compliant lap splice, the failure should, by definition, be yielding of the spliced reinforcing bars. If that's not the case, then the world's got a rather serious problem on its hands. Conversely, from a practical design perspective, do we care how non-compliant lap splices fail?

 

[asiga]

Thanks a lot for your drawing and explanation, KootK, but now I really don't understand how you calculate the "anchorage length".

A quick thought: maybe you call "development length" to what we call "basic anchorage length" (which depends on just the concrete strength, the steel yield strength, the diameter of the rebar, and the position of the rebar at the time concrete is cast).

...and then maybe your "anchorage length" is what we call "design anchorage length" (which depends on the basic anchorage length, as well as the actual stress of the rebar and the anchorage procedure: straight, bend, hook, loop, welded transverse bar,...). Obviously, our "design anchorage length" is always less or equal than the "basic anchorage length".

If your difference from the "development length" to the "anchorage length" has nothing to do with any of this, then I really don't understand how can you consider in your "anchorage length" formulas the phenomena you are describing without using a strut-tie model.

But, anyway, in your first post you seemed to say that development and anchorage were interchangeably used in North America in the past... so... maybe there was a change in the theory models at some time? When did it happen? Can you point me to a book which you would consider as the key document that clearly defined the difference?

 

[KootK]

Thanks a lot for your drawing and explanation, KootK, but now I really don't understand how you calculate the "anchorage length".

1) If it's reinforcing anchored by concrete in tension (3b above) then it's the Eligenhausen anchorage stuff often referred to as "Appendix D" in North America.

2) If it's reinforcing anchored by a concrete strut or compression field (3a above), then the anchorage checks are really built into the normal checks for shear and flexure and additional checking to verify anchorage is not required.

Feel free to post a sketch of a situation in which you're not sure how to satisfy the anchorage requirement and I'll do my best to show you how it's done.

A quick thought: maybe you call "development length" to what we call "basic anchorage length" (which depends on just the concrete strength, the steel yield strength, the diameter of the rebar, and the position of the rebar at the time concrete is cast).

That sounds promising but I'd have to see the particular provisions in order to comment on them intelligently. Can you post them or, at least, provide the names of the standards and the relevant clauses within them?

If your difference from the "development length" to the "anchorage length" has nothing to do with any of this, then I really don't understand how can you consider in your "anchorage length" formulas the phenomena you are describing without using a strut-tie model.

The sketch below shows just such an example and how it is dealt with without resorting to strut and tie. Effectively, adequate shear and flexure design in that case provides the necessary anchorage capacity. Of course, our provisions for shear design when using stirrups is really utilizing a strut and tie truss model so, in that sense, it is still related to strut and tie.

But, anyway, in your first post you seemed to say that development and anchorage were interchangeably used in North America in the past... so... maybe there was a change in the theory models at some time?

No change. The requirement for reinforcing steel to be anchored by something, most often concrete compression struts/fields, has been a requirement of reinforced concrete design since it's inception. I think that the confusion comes about because:

1) Authors consider this requirement to be so fundamental to concrete design methodology that it doesn't warrant being stated explicitly.

2) Because type 3a anchorage is satisfied indirectly via conventional shear an moment checks, designers don't realize that these checks are, in part, addressing the anchorage requirement.